RESEARCH

Since the summer of 2015, I have been a research assistant in the Dept of Systems and Information Engineering, working on a variety of projects within the health sciences. A summary of my current projects is provided below.
   

LINKS FOR CURRENT PROJECTS


Predictive Modeling of Sepsis - abstract here for project presented at 2017 BMES National Conference
Dept of Systems and Information Engineering & Dept of Medicine, University of Virginia

Trauma Acuity Stratification - manuscript here under review for publication by Annals of Emergency Medicine
Dept of Systems and Information Engineering & Dept of Surgery, University of Virginia

Computer-Aided Prescribing in Geriatrics - project details here
Dept of Systems and Information Engineering & Dept of Medicine, University of Virginia

Wnt/β-Catenin Pathway in Xenopus laevis - contact me if interested in project details
Dept of Biology, University of Virginia

Patient-Centric Design for T1D Self-Care - preparing manuscript here for submission to Applied Clinical Informatics
Dept of Systems and Information Engineering & Dept of Public Health, University of Virginia


PREDICTIVE MODELING OF SEPSIS

Dept of Systems and Information Engineering & Dept of Medicine, University of Virginia
_________________________________________________________________________________________________________________________

                  

 

Current Project: May, 2017 - Present

Predictive modeling of sepsis in adult ICU patients. Philip H. Schroeder, Roman Wang, Catherine Sun, Yasasvini Puligundla, Mawulolo Ameko, Christopher C. Moore, Laura E. Barnes. August, 2017.

 

Abstract of oral presentation given at 2017 Biomedical Engineering Society National Conference, Phoenix, AZ.

 

Overview:

I work with the UVa Division of Infectious Disease and International Health to build predictive models of sepsis with the goal of improving the precision and timeliness with which infections can be identified and treated. The models use patient demographics along with vital signs and blood culture results (pulled from a time window of 4 - 72hrs prior to infection onset) to calculate risk scores for developing sepsis in ICU patients. We test a variety of feature selection methods (e.g., LASSO, RFE, forward selection) and computational models (logistic regression, random forest, BMA, PCA, SVM) using the Sepsis-3 definition. Ultimately, we strive to integrate the algorithms within the UVa EHR system, allowing for more proficient treatment of sepsis than that achieved with the current approach of using SIRS criteria.

My work with sepsis has been tremendously rewarding, leading to strong relationships with physicians and researchers in the Dept of Medicine and to opportunities such as the NSF Mutli-Scale Systems Bioengineering REU with UVa Biomedical Engineering. Further, the use of machine learning in evaluating patients has redefined my understanding of what is possible in treating disease. The potential is seemingly limitless. With sepsis, we are currently exploring exciting new avenues, such as feeding models with datasets of written physician notes and past medical histories. I have been inspired by the power of these algorithms and am excited to be a part of their growing integration into patient care.

 




TRAUMA ACUITY STRATIFICATION

Dept of Systems and Information Engineering & Dept of Surgery, University of Virginia
_________________________________________________________________________________________________________________________

                  

 

Current Project: July, 2016 - Present

Relative Mortality Analysis of the “golden hour”: A comprehensive acuity stratification approach to address disagreement in current literature. Philip H. Schroeder, Nicholas J. Napoli, William F. Barnhardt, Madeline E. Kotoriy, Jeffrey S. Young. December, 2017.
Manuscript under review for publication by Annals of Emergency Medicine.
 

Overview:

I work with the UVa Trauma Center in developing novel acuity stratification approaches to overcome the limitations of common methodology used in analyses of pre-hospital and trauma care. For our most recent project, we employed a statistical method capable of
evaluating the impact of pre-hospital time on trauma patient outcome across the full spectrum of acuity. In doing so, we revealed the consequences of previous studies' reliance on an acuity threshold, defined a priori, to account for acuity in their patient populations. By overcoming this limitation, we revealed new insights on how pre-hospital time impacts trauma patient mortality and, perhaps more importantly, how assumptions made by traditional methodology can render analyses vulnerable to misleading results. Going forward, we are focused on evaluating larger datasets and different patient populations to improve the precision, and assess the robustness, of insights from this paper. In addition, I am working to use a machine learning approach that was built at MIT, titled Supersparse Linear Integer Modeling (SLIM), to develop a scoring system that can be used by pre-hospital providers with trauma patients in the field to quickly assess acuity. Through this work, I have grown a strong appreciation for the value of thinking critically about the assumptions made by statistical approaches and using this understanding to properly calibrate the context in which results should be interpreted.





COMPUTER-AIDED PRESCRIBING IN GERIATRICS

Dept of Systems and Information Engineering & Dept of Medicine, University of Virginia
________________________________________________________________________________________________________________________

 


Current Project: December, 2016 - Present

IT-based approach to increase utility of Beers’ and STOPP-START Criteria in prescribing geriatric patient medications. Mathew M. Kerwin, Philip H. Schroeder, Sean B. Sequeira. Started December, 2016.
 

Overview:

Beers' Criteria and STOPP-START Criteria have been developed to minimize potentially inappropriate prescribing (PIP) in geriatric medicine. However, inefficiencies in usability greatly limit their utility (i.e., they are cumbersome and tedious to reference). We sought to address this problem by building an IT system that allows physicians to paste a list of medications into a search bar that then returns the associated recommendations from both Beers’ and STOPP-START Criteria. The system is integrated with the UVA EHR System and can use patient data to alert physicians of PIP. The system is being introduced as a new tool within UVA Internal Medicine in a prospective study to test whether physicians with access to the system will more frequently reference the criteria when prescribing medications to geriatric patients than those without access. In essence, the system functions to increase the utility of the criteria by increasing their usability.




WNT/β-CATENIN PATHWAY, XENOPUS LAEVIS

Dept of Systems and Information Engineering & Dept of Medicine, University of Virginia
________________________________________________________________________________________________________________________

                
 

Current Project: July, 2017 - Present

Effect of temporal perturbation of Wnt/β-catenin pathway with IWP-2 on Xenopus morphological development. Nayla Labban, Melissa M. Le, Claire M. Romaine, Philip H. Schroeder, Natasha M. Smook, Raymond E. Keller. Started July, 2017.
 
Contact me if interested in project details.
 

Overview:

The Wnt/β-catenin pathway is conserved throughout metazoans and is linked to a variety of diseases, including multiple cancers, bone disorders, and metabolic disorders. In this study, we seek to assess how this pathway is associated with morphological development in Xenopus laevis. Morphology and β-catenin behavior will be assessed following treatment with IWP-2 at different developmental stages.




 

T1D PATIENT-CENTRIC SYSTEM DESIGN

Dept of Systems and Information Engineering & Dept of Public Health, University of Virginia
________________________________________________________________________________________________________________________

                         

 
Current Project: September, 2017 - Present
:

Exploring psychosocial challenges in type-1 diabetes self-management during transition to college as a foundation for informed design of personalized technological aids. Philip H. Schroeder, Dylan A. Hazlett, Kelly E. McCain, Syeda Z. Narmeen, Taylor K. Puhl, Rupa S. Valdez. Started September, 2017.
 
Preparing manuscript for submission to Applied Clinical Informatics.
 

Overview:

In this study, we conducted semi-structured interviews of U.S. college students, ages 18-24, with type-1 diabetes (T1D) to investigate the psychosocial challenges of self-management during the transition to college and the potential for designing new devices to aid with such challenges. Interviews were transcribed and coded using inductive qualitative content analysis. Insights gained through the interaction with study participants were combined with existing literature on T1D self-care and relevant theoretical frameworks to provide a guiding foundation for future development of technological aids.






   
 

_________________________________________________________________________________
_________________________________________________________________________________

phs5eg AT virginia.edu
(434) 589-5935